Electrical conductivity is an important property for technological applications of nanofluids that has not been widely studied. Conventional descriptions such as the Maxwell model do not account for surface charge effects that play an important role in electrical conductivity, particularly at higher nanoparticle volume fractions. Here, we perform electrical characterizations of propylene glycol-based ZnO nanofluids with volume fractions as high as 7%, measuring up to a 100-fold increase in electrical conductivity over the base fluid. We observe a large increase in electrical conductivity with increasing volume fraction and decreasing particle size as well as a leveling off of the increase at high volume fractions. These experimental trends are shown to be consistent with an electrical conductivity model previously developed for colloidal suspensions in salt-free media. In particular, the leveling off of electrical conductivity at high volume fractions, which we attribute to counter-ion condensation, represents a significant departure from the "linear fit" models previously used to describe the electrical conductivity of nanofluids.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to The Review of Economics and Statistics. Abstract-It is well known that least squares estimates can be very sensitive to departures from normality. Various robust estimators such as least absolute deviations (LAD), Lp estimators or M-estimators provide possible alternatives to least squares when such departures occur. This paper applies a partially adaptive technique to estimate the parameters of Sharpe's market model. This methodology is based on a generalized t-distribution and includes as special cases least squares, LAD, Lp as well as some estimation procedures which have bounded and redescending influence functions.

Previous heat transfer studies of nanofluids have shown that suspended nanoparticles can affect thermal properties within a fluid and furthermore can affect surface roughness by depositing on a heater surface. Pool boiling studies of nanofluids have demonstrated either enhanced or diminished heat transfer, yet have been unable to distinguish the contributions of increased surface roughness and suppression of bubble transport by suspended particles because they have used base fluids on a clean boiling surface as a comparison. We resolve this uncertainty by studying the boiling performance of a surface exposed to a series of boiling tests that alternate between water and a water-based nanofluid with suspended 40 nm ZnO nanoparticles. We find that the performance for the water tests increases significantly, showing a 62% enhancement after four cycles. This increase correlates well with a surface roughness model for boiling that uses atomic force microscopy-measured surface data to quantify the layering of nanoparticles in intervening nanofluid boiling tests. We find that the performance of the ZnO nanofluid initially shows a 24% enhancement versus water on a clean (unroughened) surface, but then steadily declines in later tests as nanoparticle layering occurs, showing a measured trend that is opposite that of water. We ascribe this decrease to the suppression of bubble formation and motion by the suspended particles. The results demonstrate that the effect of increased surface roughness due to nanoparticle layering can be twofold, greatly enhancing boiling for the base fluid and slightly decreasing performance for the nanofluid.

Kky Worcis and Phrases: regression models; adaptive mmaXlmum likelihood estimation; generalized methad of m o m e m estimmion; M-estimation; partiaIly adaptive esriman'on; m o m carb simuIation JEL classification: C13, C14 ABSTRACT Numerous estimation techniques for regression models have beenproposed. These procedures differ in how sample information is used in the estimation procedure. The efficiency of least squares (OLS) estimators implicitly assumes normally distributed residuals and is very sensitive to departures from normality, particularly to "olrtliers" and thick-tailed distributions. T , &US,--:e deviation (LAD) estimators are less sensitive to outliers and are optimal for Laplace random disturbances, but not for normal errors. This paper reports Monte Carlo comparisons of OLS, LAD, two robust estimators discussed by Huber, three partially adaptive estimators, Neweyys generalized method of moments estimator, and an adaptive maximum likelihood estimator based on a normal kernel studied byManski. This paper is the f i r s t to compare the relative performance of some adaptive robust estimators (partially adaptive-and adaptive procedures) with some common nonadaptive robust estimators. The partially adaptive estimators are based on three flexible parametric distributions for the errors.These include the power exponential (Box-Tiao) and generalized t distributions, as well as a distribution for the errors, which is not necessarily symmetric. The adaptive procedures are "fully iterative" rather than one step estimators. The adaptive estimators have desirable large sample properties, but these properties do not necessarily carry over to the small sample case.The Monte Carlo comparisons of the alternative estimators are based on four different specifications for the error distribution: a normal, a mixture of normals (or variance-contaminated normal), a bimodal mixture of normals, and a lognormal. Five hundred samples of 50 are used. The adaptive and partially adaptive estimators perform very well relative to the other estimation procedures considered, and preliminary results suggest that in some important cases they can perform much better than OLS with 50 to 80% reductions in standard errors.